Investigating the Effect of Flow and Sediment Particles Characteristics on Sandy Sediments Transport in Circular Sections using Data Driven Methods

Sedimentation is one of the important phenomena in water engineering due to its effect on the transport capacity and hydraulic performance of water transport structures. As there are numerous factors that affect this phenomenon, it is difficult to accurately determine the most influential parameters. In the present study, the efficiencies of intelligent Gaussian Process Regression (GPR) and Adaptive Neuro-Fuzzy Inference System (ANFIS) approaches in the prediction of sandy sediments in circular pipes with rough and smooth beds have been evaluated. Using several series of laboratory data, different models were defined with considering the impacts of hydraulic parameters and sediment particle characteristics and evaluated for rough and smooth pipes. The results showed a high accuracy of the methods used in the present research. According to the results in estimating the sediment load in circular pipes using only hydraulic parameters did not lead to accurate results, and the properties of sediment particles also affect the estimation process of this parameter. With performing sensitivity analysis, it was observed that the particle's Froude number is the most important parameter in the estimation of sediment load in circular pipes. Also, the results showed that the bed and walls roughness of pipes was effective in sediment transport and the increase of roughness values reduced the accuracy of the results.